Roughly two document reading techniques are available for image reading apparatuses such as digital copiers, scanners, and facsimile machines. One of the two techniques is to read an image from a document by moving an optical system with the document placed on a original plate to fix the position of the document (optical system moving method). The other is to read an image from a document by using an auto document feeder (ADF) to feed the document, with the position of the optical system fixed (document moving method).
In particular, with the document moving method, rubbish attached to the original plate may disadvantageously appear in the document image as a line-like image.
To solve this problem, a method has been proposed (Japanese Patent Laid-Open No. 2000-196814). This method is used in a sheet through document reading apparatus having a document feeder. When the apparatus performs a main scanning operation at a predetermined time and detects an abnormal data value, a document-read position is changed in accordance with a predetermined method.
A method for detecting rubbish on the basis of a read image has also been proposed (Japanese Patent Laid-Open No. 2000-152008). With this method, if any portion in an image read by one of plural reading means does not match the same portion in an image read by another reading means, this portion is determined to be a line-like image.
However, the invention described in Japanese Patent Laid-Open No. 2000-196814 detects, as a line-like image, consecutive image data each having a value equal to or larger than a predetermine value or equal to or smaller than a predetermined value. Consequently, even image data for a vertical line present in a document is mistakenly detected as rubbish.
Further, the invention described in Japanese Patent Laid-Open No. 2000-152008 can detect rubbish attached only to the document-read position of one of the plural reading means. However, this technique is disadvantageous in that rubbish such as waste yarn or hair which moves with the document cannot be detected. This is because such rubbish moves together with the document between the document-read positions of the corresponding reading means, so that images read at these document-read positions match.